MarketMaker techniques

[FredBloggs]

Quote from StillStanding:

Hi

How do big firm market makers handle trend risk?

If always maintining bid/ask spread , then a rapid trend move could result in them averaging down all day against the trend ending up with enormous losses at end of day.

How do they avoid this?

Do they carry over the loss?

Do they stop averaging down at some point?

Do they bail out at some point? if so when?

Do they keep moving the average down point further away as their losing position grows?

Any former professional market makers working for a big firm on this forum ?

Thanks

as youve probably gathered, directional trading (trends) isnt the only way to make money.

neither are mm averse to taking losses. many exchanges often have liquidity provider programms for members to get rebates for quoting/providing liq for new contracts. they often act as mm quoting both sides of a market, hedging in a related contract, perhaps on a different exchange. there has been quite a lot of this in recent years in commodity markets. the result is they get massively stung as they find themselves on the wrong side, limit up in an illiquid contract they cant get out of, while their 'hedge' isn't doing much at all.

mm aint the free ride it would appear to be.

 

[TokyoGhetto]

I have been playing with a market making strategy for some time now. Below you will find attached a sample of my trading.

My observation is that even with a razor thin PnL a market maker can be profitable depending on the amount of share they can execute (market share)

for example my BAC sample has over 400 executions and the difference in cost avg price is .001. On 21000 shares it resulted in a Gross Profit of $23.23. Of course this model crossed market so it resulted in a Net Loss after fees.

My UYG sample was able to break even and be net positive after fees due to the rebates. this model only crossed market to fill up on inventory at the beginning and crossed on losing positions.

My CIT sample was a volatility sample (trading was done when bankruptcy was announced) The model managed to break even and might have been net positive if the stock was trading over a dollar after rebates.

It is a game of averages and HFT models need to recalibrate all the time. My goal would be to maintain a positive gross whilst trading as much volume as possible and shut the model down once it reaches break even.

.001 on a million shares is $ 1000
.0025 rebate on a million shares (add liq) = $ 2500

so if the model can trade a million shares and shut down within .001 of a cent before going negative, then its a nice payday.

If i can do this and achieve a gross greater then .02 on a million shares i'm co-locating.

BTW BAC does over 100 million shares a day.

At this point i ask myself, whats the next step. maybe apply the model to high alpha stocks, so that the model has time to churn as many shares as it can before the gross hits break even.

Anyone have any input on this?

 

[TokyoGhetto]

just found this on zerohedge....good read on HFT MM.

"
The one factor consistent across all HFT strategies is that they benefit from increased volumes and micro-second execution advantages. The majority of profits are made on razor thin margins. For example, market makers are often executing trades with gross margins of 0.05% or nominally between one and two-tenths of a penny. Of course most of the volume in equity markets is concentrated in the largest capitalization stocks with the narrowest spreads and the lowest price volatility. These characteristics attract the attention of HFT strategies; essentially, trading begets trading in these names. Non-HFT trades (now thought to be less than 50% of overall equity trading) are necessary for HFT strategies to flourish. In order to manage the trading velocity required by HFT strategies, firms must often submit and cancel thousands of orders per second while simultaneously monitoring the market data that drives their automated trading systems. To get the broadest and fastest access, many HFT firms subscribe to enhanced market center data feeds like NASDAQ’s ITCH or BATS’ FASTPITCH, in place of feeds from the slower consolidated Securities Information Processor (SIP). In addition to the expensive data feeds, these firms have to invest heavily in their internal systems and data processing engines in order to monitor and process data on thousands of securities simultaneously. [so yeah, all it takes is an i7 and a $29.95/month program to run a HFT system, right]. The time frames under consideration are so short that many of these firms have taken the step to co-locate their technology inside the buildings housing the exchanges."

http://www.zerohedge.com/article/to...duced-trading-toll-why-vwap-gold-mine-not-you

 

[Occam]

Quote from TokyoGhetto:

I have been playing with a market making strategy for some time now. Below you will find attached a sample of my trading.

My observation is that even with a razor thin PnL a market maker can be profitable depending on the amount of share they can execute (market share)

for example my BAC sample has over 400 executions and the difference in cost avg price is .001. On 21000 shares it resulted in a Gross Profit of $23.23. Of course this model crossed market so it resulted in a Net Loss after fees.

My UYG sample was able to break even and be net positive after fees due to the rebates. this model only crossed market to fill up on inventory at the beginning and crossed on losing positions.

My CIT sample was a volatility sample (trading was done when bankruptcy was announced) The model managed to break even and might have been net positive if the stock was trading over a dollar after rebates.

It is a game of averages and HFT models need to recalibrate all the time. My goal would be to maintain a positive gross whilst trading as much volume as possible and shut the model down once it reaches break even.

.001 on a million shares is $ 1000
.0025 rebate on a million shares (add liq) = $ 2500

so if the model can trade a million shares and shut down within .001 of a cent before going negative, then its a nice payday.

If i can do this and achieve a gross greater then .02 on a million shares i'm co-locating.

BTW BAC does over 100 million shares a day.

At this point i ask myself, whats the next step. maybe apply the model to high alpha stocks, so that the model has time to churn as many shares as it can before the gross hits break even.

Anyone have any input on this?

TokyoGhetto,

Not sure what those results are -- is that some sort of backtesting system? If so, you might find that reality yields profoundly different results. A lot of order flow is internalized these days, and unless your backtester is taking this into account, you'll find that much of the volume thus simulated is in fact not accessible to the open markets, but rather is going to dark pools or is internalized to brokerages, which means that the B/D is trading against their own clients at their discretion.

 

[Occam]

Quote from TokyoGhetto:

just found this on zerohedge....good read on HFT MM.

I would take whatever you find on zerohedge with a grain of salt. It seems the author has an ax to grind and merely cherry-picks studies/articles/quotes that support opinions he already has.

 

[TokyoGhetto]

My data is from live execution. 100 shares per execution. I did it on cheap, liquid stocks so I wouldn't get burned on an outlier movement.

Wish I applied the model on AA (alcoa) or SNDK (sandisk). The increase in volitility would help the HFT model trade through more shares before reaching the breakeven point. I would still need to decide in what direction to trade the model in (long/short).

Seems like everyone is doing it on stocks like BAC and what ever edge was present is thin at this point unless option traders can push the vix to 60 again and wake up the beast.

Ohh and I do take what zero hedge provides with a grain of salt I only take verified research data seriously. All the conspiracy stuff gets thrown out the window.

 

[zdreg]

Quote from el pollo:

I never knew "market makers" in today's market took positions!

NYSE specialists don't "maintain" an orderly market, and NASDAQ MM'ers certainly don't these days, well they run ahead of the customer, if you call that market making!

The good old days of flipping CSCO for 1/2 points are long over.

I think I saw a picture of a NYSE specialist taking a position in the Luncheon Club!

Which market are you referring to?

"The good old days of flipping CSCO for 1/2 points are long over. "

nothing is forever. perhaps with a transaction tax u will get your wish.

u know the old saying: don't wish for something u may receive it.

 

[luckybastard]

Quote from def:

Out of curiosity, how do you measure the "dumb ones vs the very smart ones" as profitability doesn't seem to be your main criteria?

Dumb = you don't need any brains to do the job: HO does the thinking for you. Smart = you need to pass the most difficult screening tests in the industry to get the job and you'll need to use your brains to get your job done. And no, that doesnt necessarily translate directly into profit, I've never said that.

 

[Occam]

Quote from TokyoGhetto:


Seems like everyone is doing it on stocks like BAC and what ever edge was present is thin at this point

I agree 100% (although my impression is that there's also dramatically increased competition even in the smaller-volume stocks in recent months).

Regarding zerohedge "HFT bugbear" Goldman, I've read in several analyses that most of Goldman's trading profits come from OTC/illiquid/derivative instrument trading, not equity HFT.

 

[nitro]

Quote from TokyoGhetto:

...

Anyone have any input on this?

I don't see how anyone, at least I can't, comment unless you:

a) Show what your cost structure is so that we can see where the $23 came from. If all you have to do is reduce your costs of doing business, and there is room, that seems simple enough.

b) I don't understand what you are doing. You say you are making money from the rebates on the stock. Does that mean you are reverse hedging a stock position with options to get a rebate?